Date:         Thu, 10 Apr 1997 16:55:57 -0500
Reply-To:     Yuxian Jiang <joan.jiang.10@ND.EDU>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         Yuxian Jiang <joan.jiang.10@ND.EDU>
Subject:      Re: changing units of analysis
In-Reply-To:  <199704101710.AA20212@gate1.health.state.ny.us>
Content-Type: text/plain; charset="us-ascii"

Please modify the following codes so that it reads whole input data.  File
OUT1 holds the data with the format that you want. Good luck!    --Joan
Jiang
 
DATA LIST
   /PID   1-2  (A)        /*personid
    FID    5-7  (A)        /*familyid
    OTH 10-12 (A)        /*other stuff
BEGIN DATA
01  001  234
02  001  098
03  002  743
04  003  432
05  003  321
06  003  431
END DATA
 
SORT CASES BY FID
IF (FID NE LAG(FID,1)) X=0 /*Initialize for new FID
COMPUTE X=X+1
LEAVE X
FORMAT X (F1)
EXE
 
/*Assume that the max number of record per familyid is 9
VECTOR VPID(9A3) VFID(9A4) VOTH(9A4)
COMPUTE VPID(X)=PID
COMPUTE VFID(X)=FID
COMPUTE VOTH(X)=OTH
AGGREGATE OUTFILE=*
          /BREAK=FID
          /VPID1 TO VPID9=MAX(VPID1 TO VPID9)
          /VFID1 TO VFID9=MAX(VFID1 TO VFID9)
          /VOTH1 TO VOTH9=MAX(VOTH1 TO VOTH9)
WRITE OUTFILE=OUT1
      /VPID1 VFID1 VOTH1 VPID2 VFID2 VOTH2 VPID3 VFID3 VOTH3
         VPID4 VFID4 VOTH4 VPID5 VFID5 VOTH5 VPID6 VFID6 VOTH6
         VPID7 VFID7 VOTH7 VPID8 VFID8 VOTH8 VPID9 VFID9 VOTH9
EXE
 
 
>Hi,
>
>We are using part of CPS data, and trying to put individuals together
>into families.
>
>That is, each individual in the CPS data set has one record, which
>includes a family ID  (called sequence number).  There are other
>individuals in the data set with the same family ID, so they are
>in the same family.  However, each family is made up of different
>numbers of individuals.  Is there any way to manipulate the data set
>to put the individuals together into a single record, one for each
>family?
>
>Currently, the data set looks something like:
>
>personid, familyid, other stuff
>
>01  001  234
>02  001  098
>03  002  743
>04  003  432
>05  003  321
>06  003  431
>
>and we would like it to look like:
>
>01  001  234  01  001  098
>03  002  743
>04  003  432  05  003  321  06  003  431
>
>so that each person with the same family id is on the same card.
>Can spss do this?  We are using spss for windows, but I presume
>it would have similar commands to spssx.
>